The Business of Palliative Medicine: Management Metrics for an Acute-Care Inpatient Unit
American Journal of Hospice & Palliative Care
Volume 18, Number 1
Meller P. Davis, MD, FCCP, Declan Walsh, FACP, Kristine Nelson, MD, Dale Konrad, MBA, Susan B. LeGrand, MD
The Center to Advance Palliative Care - The Business of Palliative Medicine
For any palliative medicine inpatient unit to be economically viable, certain management metrics need to be followed. Palliative medicine can provide both a compassionate and economical service within the current acute inpatient hospital environment. In this article, we will review the administrative and financial factors we have identified that influence the business of acute palliative medicine.
Keywords: acute-care units, business metrics, case-mix index, diagnosis-related groups, inpatient unit, medical economics, palliative care, palliative care program, palliative medicine, palliative medical department
Inadequate care of patients and the families of patients with advancing chronic illness in hospitals is well documented. Twelve percent or more of patients in acute hospital wards have advancing chronic disease and are appropriate for palliative medicine services.1, 2 These patients may have metastatic cancer, severe chronic obstructive lung disease, multiple organ failure or HIV/AIDS. Pain is present in 39 percent, weakness in 33 percent, constipation in 31 percent, agitation in nearly 25 percent and dyspnea in 22 percent.1, 2 Nearly one-quarter have worries or concerns beside their medical illness, and almost one-third of relatives' concerns are about the patient and not directly related to his or her medical condition.1
Only 30 percent may be referred to palliative medicine services.1, 2 In one report, discharges were uncoordinated or disorganized in 27 percent.1 This resulted in extra days of hospitalization and readmissions. Ongoing community palliative services were involved in only 17 percent.1 There was poor documentation of discharge plans and sub-optimal advance care planning.
Hospital palliative medicine services and inpatient units can facilitate symptom control, management of complications, patient and family understanding and disposition with placement in a community care setting, as needed.3, 4 Palliative services are economical. The duration of hospitalization is shortened, resources are used appropriately (unnecessary testing is eliminated), and readmissions are limited due to effective disposition.5-7 However, for palliative medicine inpatient units to be developed, economic viability is necessary.8
The purpose of this article is to review factors that affect the economic viability of inpatient palliative medicine units. These factors, or metrics are based on our experience in managing a 23-bed, acute-care palliative medicine unit, since July 1994, the first of its kind in the United States. The unit admitted more than 900 patients in 1999 for symptom control and treatment of complications of advanced disease. The unit is situated in the Cleveland Clinic Foundation, a 1000-bed tertiary care teaching hospital. Systematic focus on the metrics identified will provide a template upon which a successful palliative inpatient unit may be established and maintained within an acute-care hospital in the United States. We hope this information may be helpful to other institutions starting similar programs.
Multiple factors influence the economics of inpatient units and include clinical issues such as census, length of stay, plan of care, discharge planning, family meetings, full-time equivalent employees (FTE) and formulary. Nonclinical factors that require monitoring include the direct costs of care per bed-day, indirect costs, contribution margins and net revenue, net income, diagnosis-related group (DRG) and case-mix index (CMI).
For palliative medicine units to survive, a high census is necessary. Our inpatient unit consistently has had a census of more than 80 over the past year. Economic viability is threatened with a census below 70 percent. The census has grown due to increased utilization by Cleveland Clinic consultants from the oncology department as well as cardiology, internal medicine, intensive care, neurology and various surgical department.
Length of stay
length of stay directly affects the total and indirect costs of care. Our median length of stay was 8.1 days in 1999. The cost of care will exceed net revenues as a rule after 10 days of hospitalization. The DRG is a poor predictor of length of stay. Case-mix index also may not accurately predict hospitalization due to clinical progression of cancer during hospital stays, resulting in further unpredictable loss of function.10, 11
Plan of care
The length of stay may be unnecessarily prolonged by failing to:
consideration is that a plan of care be established for symptom management and
medical care within 24 hours. This plan should be updated daily with anticipated
date of discharge during the daily interdisciplinary team meeting. A general
community-based plan of care with an established date of discharge should later
follow a family meeting. The family meeting should occur within 48 to 72 hours
of admission to explore broader issues, including advance care directives,
psychological issues, spiritual concerns and family understanding of the
patient's illness. Ninety-three percent of the patients have family meetings
during their stay on our inpatient unit.
- establish a plan of care early upon admission
- aggressively treat symptoms
- establish a family meeting
Discharge planning is key to limit the length of stay. Discharge planning with family involvement will reduce potential readmissions due to effective disposition, based on realistic family resource. This requires one full-time medical social worker for our 23 beds. Discharge planning in order to be successful requires a full-time registered nurse (RN) as discharge coordinator per 23 beds to act as liaison for community resources. A daily report is done before work rounds to update the medical condition of the patient and to review the outcome of the family meetings and goals of care, with disposition date in mind, with the social worker and discharge coordinator present. Once a week, an interdisciplinary team meeting reviews patient disposition, plans for follow-up and changes in symptom or performance status of inpatients during the prior week.
Forty percent of our acute palliative medicine patients are hospice appropriate on discharge. A close link needs to established with hospice programs to facilitate the transfer to home. The ideal arrangement consists of acute palliative specialists who also provide continuity as hospice physicians. Patients can deteriorate during their stay and thus exceed the length of stay predicted upon admission. Advance directives are established in 91 percent of our admissions. Twenty percent of patients die whole on the unit, and if a patient is actively dying as an inpatient, he or she is cared for until death and not discharged. Economic consideration should be ignored in this situation.
Full-Time Equivalent Employees (FTEs)
A balance of nurse-to-patient ratio and adequate number of full-time nurse equivalent employees is necessary for excellent care. Nursing satisfaction without "burnout" and economic feasibility are dependent on this balance. An RN-to-patient ratio of 1:4 in the day and evening shifts, and 1:5 for the night shift, is ideal. Job satisfaction for nursing staff depends on the number of hours worked, "creative" (flexible) shifts to meet personal needs, support from the medical staff and administration and the development of outside personal interests. Lack of job satisfaction lead to rapid turnover, reduced quality of care and added stress on remaining nursing staff.
The proper number of nursing assistants (NAs) and licensed practical nurses (LPNs) is invaluable in effectively utilizing RN skills. NAs provide ancillary care without direct patient responsibility, whereas LPNs can assume direct nursing care of individual patient. One LPN and two or three NAs for five RNs appears optimal. A ratio of patient care nursing assistants (PCNAs) to patients of 1:6 is best.
Our 23-bed inpatient palliative medicine unit employs two to three physician extenders and one palliative medicine fellow, who rotates monthly, as responsible for inpatient clinical care. One consultant on a rotating basis supervises the trainees and provides educational support. On average, one full-time consultant will see 200 to 250 newly referred patients per year.
Ancillary personnel particular to the palliative medicine inpatient unit include a nurse discharge coordinator, medical social worker, music therapist, respiratory therapist (RT), occupational therapist (OT) and physical therapist (PT). Night-time coverage is shared with internal medicine residents with a fellow consultant available 24 hours.
Drug costs increase with census. A well-developed formulary of established medications and standard protocols of care will control medical costs and provide a template for prospective research and quality-of-care audits. There is a tendency to use newer, well-promoted medications, which are expensive. They should prove to be better or have specific advantages compared to standard medications before being adopted. Medications should be prescribed by generic rather than brand name. Brand names differ from country to country and limit communication and education.
Direct cost of care
Direct cost of care per bed-day is determined by nursing costs and pharmacy costs and is inversely proportional to census. Medical nursing costs account for 55 percent to 62 percent of direct costs, while pharmacy costs constitute 20 percent. A high census reduces direct costs by resource sharing and optimizing FTE utilization. Palliative inpatient units do laboratory tests and radiographic procedures for symptom management and assessment of treatment response. Chemotherapy and radiation therapy are also option, and hence cost differ significantly from hospice inpatient units. However, laboratory, radiology and chemotherapy play a minor role in direct costs. Direct costs can be monitored by the unit and are influenced by medical decisions, unlike indirect costs. Such decisions should not be to the detriment of patients.
Indirect cost of care
Indirect cost is related to the institutional (hospital) ancillary services and should be equally distributed to all units with the hospital. Our indirect costs of care are a mean 45 percent of the total cost of care, but vary by length of stay. The individual palliative medicine unit has no control over indirect costs.
The contribution margin is the net revenue minus direct costs. The contribution margin percentage is the contribution margin divided by net revenue times 100. Our mean contribution margin percentage was 48 percent in the first eight months of 1999. Contribution margin does not appear to be related to length of stay. The contribution margin does not take into account indirect costs, which are the main determinant of costs in a prolonged hospital stay.
Net revenue and net income
Net revenues will depend on insurance contracts, i.e., capitated fees versus resource utilization or DRG-based net revenues. Our net income is determined by census, duration of stay, and case severity as reflect in the case-mix index. Our net revenue, depending on individual DRG, case severity and length of stay, is a mean of 50 percent of gross revenue. Unnecessary delays by ancillary services (i.e., radiology, radiation, and medical consultants) will extend hospitalization and reduce net income and economic viability. Radiation should be in short course (preferably single-fraction) and continued through the weekend.
Diagnosis-related group (DRG)
The reimbursement by Medicare for hospital care is currently based on a diagnosis-related group. This system was developed in 1983 for short-term hospital prospective payments. The disadvantage of the DRG system is that it stratifies poorly for disease severity and need-based resource utilization, particularly in palliative medicine.9 The DRG does not predict length of stay in advancing chronic illness.9-11
Case-mix index (CMI)
The case-mix index is a diagnosis-related group with a relative weight scale based on resource, labor and supply utilization (see Table 1).
|Table 1. Case-mix index by DRG in 1999|
|Diagnosis-related group (DRG)||Total cases or visits||Case-mix index (CMI)|
|1) 082 Respiratory neoplasm||79||1.339|
|2) 239 Pathological fracture||50||0.981|
|3) 172 Digestive Malignancy||48||1.291|
|4) 403 Lymphoma and non-acute leukemia||48||1.693|
|5) 203 Malignancy of hepatobiliary tract||46||1.274|
CMI indirectly reflects severity of illness and better predicts cost of care as compared to DRG. CMI does not predict length of stay. CMI allows comparisons between units by disease severity. Our five most frequent DRGs with CMI are shown in Table 1. A palliative care case-mix index classification has been developed and uses phases of care (i.e., acute, stable, deteriorating, terminal, and bereavement) and severity of symptoms.11 A case-mix index based on functional capacity and activities of daily living applied to palliative units in Australia predicted direct bed-day costs and highly correlated with direct costs of care per day-bed at the Cleveland Clinic.12 For a CMI of 1:36, the direct mean cost per day-bed is $248 (Australian), which is remarkably close to our direct mean cost per day-bed in American dollars. However, in light of the exchange rate, the mean cost per day-bed is 50 percent higher in the United States.
Administrative, clinical and financial metrics have been identified for the management of an acute-care inpatient palliative medicine service. Acute palliative medicine is economically feasible within the modern hospital environment, provided that one identifies and adheres to the management metrics. The commonly used DRG does not predict length of stay, case severity or resource utilization of palliative medicine inpatient units. CMI is predictive of resource consumption and cost of care, but is less predictive of length of stay. Direct cost can be influenced by medical decisions, such as FTEs, formulary choices, aggressive symptom management, established plans of care, family meetings, and direct links to hospice, which shorten the length of stay, a major contributor to economic viability. A palliative medicine case-mix index would help compare demographics, case severity and direct costs per CMI, nationally and internationally.
Mellar P. Davis, MD, FCCP, Medical Director, Harry R Horvitz Center for Palliative Medicine (a World Health Organization Project); Cleveland Clinic Taussig Cancer Center, Cleveland, Ohio.
Decian Walsh, MSc, FACP, Harry R Horvitz Chair in Palliative Medicine; Harry R Horvitz Center for Palliative Medicine (a World Health Organization Project); Cleveland Clinic Taussig Cancer Center, Cleveland, Ohio.
Kristine Nelson, MD, Lee and Jerome Burkons Scholar in Palliative Medicine; Harry R Horvitz Center for Palliative Medicine (a World Health Organization Project); Cleveland Clinic Taussig Cancer Center, Cleveland, Ohio.
Dale Konrad, MBA, Director, Special Projects, Division of Finance; Cleveland Clinic Foundation, Cleveland, Ohio.
Susan B. LeGrand, MD, Harry R Horvitz Center for Palliative Medicine (a World Health Organization Project); Cleveland Clinic Taussig Cancer Center, Cleveland, Ohio.
- Edmonds P, Karlsen S, Addington-Hall J: Palliative care needs of hospital inpatients. Palliat Med.
2000; [in press].
- Morize V, Nguyen DT, Lorente C, Desfosses G: Descriptive epidemiological survey on a given day in all palliative care patients hospitalized in a French university hospital. Palliat Med. 1999; 13:105-117.
- Ellershaw JE, Peat SJ, Boys LC: Assessing the effectiveness of a hospital palliative care team. Palliat Med. 1995; 9:145-152.
- Edmonds PM, Stuttatford JM, Penny J, et al.; Do hospital palliative care teams improve symptom control? Use of modified STAS as an evaluation tool. Palliat Med. 1998; 12:345-351.
- Axelsson B, Christensen SB: Evaluation of hospital-based palliative support service with particular regard to financial outcome measures. Palliat Med. 1998; 12:41-49
- Witteveen PO, van Groenestijn MAC, Blijham GH, Schrijvers APJ: Use of resources and costs of palliative care with parental fluids and analgesics in the home setting for patients with end-stage cancer. Ann Oncol. 1999; 10:161-165
- Lonberger EA, Russell CL, Burton SM: The effects of palliative care on patient charges. JONA. 1997;
- Walsh D, Gombeski WR, Goldstein P, et al.; Managing a palliative oncology program: the role of a business plan. J Pain Symptom & Manage. 1994; 9(2):109.
- Rutledge R, Osler T: The ICD-9 based illness severity score: a new model that out performs both DRG APR-DRG as predictors of survival and resource utilization. J Trauma.
- Davis MP, Walsh D, Nelson K, et al.; Palliative medicine: the economics of acute
inpatient care [Unpublished manuscript, in press].
- Eager K, Cromwell D, Kennedy C, Lee L: Classifying sub-acute and non-acute patients: results of the New South Wales case-mix area network study. Aust Health Rev. 1997; 20:26-42.
- Lee L, Kennedy C, Aitken J: Does the NAIP classification predict length of stay in rehabilitation, geriatrics and palliative care? Aust Health Rev. 1996;
- Goldstein P, Walsh D, Horvitz LU: The Cleveland Clinic Foundation Harry R Horvitz Palliative Care Center. Support Care Cancer. 1996; 4:329-333.
- Walsh D: Continuing care in a medical center: The Cleveland Clinic Foundation Palliative Care Service. J Pain Symptom & Manage. 1990; 5(5):273-277.
- Tropiano P, Walsh D: Organization of services and nursing care: hospice and palliative medicine. Semin Oncol. 2000; 27:7-13.
Managing a Palliative Oncology Program: The Role of a Business Plan
Journal of Pain & Symptom Management
Volume 9, Number 2
Other Resources By
Meller P. Davis, MD, FCCP
Declan Walsh, FACP
Kristine Nelson, MD
Dale Konrad, MBA
Susan B. LeGrand, MD